I know people are probably bored of ChatGPT mistakes - but its knowledge of #DoctorWho is impressively bad!
-
I know people are probably bored of ChatGPT mistakes - but its knowledge of #DoctorWho is impressively bad!
I asked it which Oscar Winners had appeared in Doctor Who.
The first two are bang on. Number 3… eh! I'll take it.
But then it gets worse (cont.)
-
The AI has picked up the fact that John Hurt was honoured in the Academy Awards' "In Memorandum" section. Sadly, he was never a winner.
Ben Kingsley was *rumoured* to be playing Davros back in 2007 - but it never happened.
Ecclesdoc *was* in The Others. It won many awards. But not a single Oscar.
-
Look, we all love Lynda Baron - and she was excellent in both Enlightenment and Closing time.
I was surprised to find out she was in Yentl - but indeed she was!
However the Oscar went to Michel Legrand and Alan & Marilyn Bergman. Not her.What does this tell us?
-
This was GPT-4o. It excels at producing authoritative *sounding* text.
But it is utter rubbish. Its accuracy is 2.5/7. Even if you're being generous, it is wrong over half the time.
Surely the AI bubble has to burst soon?
ChatGPT
ChatGPT helps you get answers, find inspiration and be more productive. It is free to use and easy to try. Just ask and ChatGPT can help with writing, learning, brainstorming and more.
(chatgpt.com)
-
Florian Idelbergerreplied to Terence Eden last edited by
@Edent I mean there is so much money in it… might just be it kinda deflates like crypto currencies
-
@Edent I don't know what the actual answer is, but this is what ChatGPT Search returns...
-
@ianbicking again, it is 50% wrong.
-
@Edent I don't know what state the bubble is in, but getting the specific and accurate data is just not really the use case for current generation models. It's a shame that this is hardly communicated.
It can be vastly improved for specific knowledge domans with RAG and other buzzwords, but even that does not fully solves the problem.You don't go to chatgpt or whatever the current AI thing is for straight up facts, at least not yet, just like you don't take the very first result you find on Google as a fact. Well, hopefully.
If they teach this thing to automatically go to *actual*(not bing) search engines, evaluate the reputation of sources, and double-check the data, we can get some significant improvements here.
-
@Edent in fact, there is a thing that kinda does that, called Perplexity. Should be much better suited for this, as far as I know.
I tried your question, there is a result if you care to take a look (I don't know if it's actually any better) https://www.perplexity.ai/search/which-oscar-winners-have-appea-GvT78HkiRiSUIIftuGoLag#0 -
@lnkr_
It gets Capaldi's film wrong. It lists a nominee as a winner - and then refers to three winners.
Just pathetic!